"Robust chronologies" or "Bayesian illusion"? : some critical remarks on the use of chronological modelling